A developmental study of random generation in children

Random generation is a complex task, and is affected by several aspects ofresponse configuration. For example, among adult subjects, the number of responsealternatives affects response sequences; a larger set size leads to response alternativesbeing sampled with uneven probabilities. This effect has been attributed to therequirement to maintain an internal representation of all response choicessimultaneously, based on the finding that the visual provision of response alternativesremoves the set size effect in random number generation (Towse, 1998). It is ofinterest, then, to establish whether children are also affected by changes in the numberof response alternatives.